Merge f70619bfe5 into 4d05f84325

* no space after @ in decorator

test/test_jsinterp.py

@@ -577,9 +577,11 @@ class TestJSInterpreter(unittest.TestCase):
     def test_unary_operators(self):
         jsi = JSInterpreter('function f(){return 2  -  - - 2;}')
         self.assertEqual(jsi.call_function('f'), 0)
-        # fails
-        # jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
-        # self.assertEqual(jsi.call_function('f'), 0)
+        jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
+        self.assertEqual(jsi.call_function('f'), 0)
+        # https://github.com/ytdl-org/youtube-dl/issues/32815
+        jsi = JSInterpreter('function f(){return 0  - 7 * - 6;}')
+        self.assertEqual(jsi.call_function('f'), 42)
 
     """ # fails so far
     def test_packed(self):

test/test_youtube_signature.py

@@ -158,6 +158,10 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/b7910ca8/player_ias.vflset/en_US/base.js',
         '_hXMCwMt9qE310D', 'LoZMgkkofRMCZQ',
     ),
+    (
+        'https://www.youtube.com/s/player/590f65a6/player_ias.vflset/en_US/base.js',
+        '1tm7-g_A9zsI8_Lay_', 'xI4Vem4Put_rOg',
+    ),
 ]
 
 

youtube_dl/downloader/__init__.py

@@ -64,6 +64,9 @@ def _get_suitable_downloader(info_dict, params={}):
     if protocol == 'm3u8_native' and params.get('hls_prefer_native') is False:
         return FFmpegFD
 
+    if params.get('_clip_args'):
+        return FFmpegFD
+
     return PROTOCOL_MAP.get(protocol, HttpFD)
 
 

youtube_dl/downloader/external.py

@@ -469,6 +469,10 @@ class FFmpegFD(ExternalFD):
             elif isinstance(conn, compat_str):
                 args += ['-rtmp_conn', conn]
 
+        ss = self.params.get('_clip_args', '').split(',')
+        if len(ss) == 2:
+            args += ['-ss', ss[0], '-to', ss[1]]
+
         args += ['-i', url, '-c', 'copy']
 
         if self.params.get('test', False):

youtube_dl/extractor/common.py

@@ -3033,7 +3033,6 @@ class InfoExtractor(object):
             transform_source=transform_source, default=None)
 
     def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
-
         # allow passing `transform_source` through to _find_jwplayer_data()
         transform_source = kwargs.pop('transform_source', None)
         kwfind = compat_kwargs({'transform_source': transform_source}) if transform_source else {}

youtube_dl/extractor/palcomp3.py

@@ -8,7 +8,7 @@ from ..compat import compat_str
 from ..utils import (
     int_or_none,
     str_or_none,
-    try_get,
+    traverse_obj,
 )
 
 
@@ -109,7 +109,7 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
     }
     name'''
 
-    @ classmethod
+    @classmethod
     def suitable(cls, url):
         return False if re.match(PalcoMP3IE._VALID_URL, url) else super(PalcoMP3ArtistIE, cls).suitable(url)
 
@@ -118,7 +118,8 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
         artist = self._call_api(artist_slug, self._ARTIST_FIELDS_TMPL)['artist']
 
         def entries():
-            for music in (try_get(artist, lambda x: x['musics']['nodes'], list) or []):
+            for music in traverse_obj(artist, (
+                    'musics', 'nodes', lambda _, m: m['musicID'])):
                 yield self._parse_music(music)
 
         return self.playlist_result(
@@ -137,7 +138,7 @@ class PalcoMP3VideoIE(PalcoMP3BaseIE):
             'title': 'Maiara e Maraisa - Você Faz Falta Aqui - DVD Ao Vivo Em Campo Grande',
             'description': 'md5:7043342c09a224598e93546e98e49282',
             'upload_date': '20161107',
-            'uploader_id': 'maiaramaraisaoficial',
+            'uploader_id': '@maiaramaraisaoficial',
             'uploader': 'Maiara e Maraisa',
         }
     }]

youtube_dl/extractor/youtube.py

@@ -2315,8 +2315,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             player_response,
             lambda x: x['captions']['playerCaptionsTracklistRenderer'], dict)
         if pctr:
-            def process_language(container, base_url, lang_code, query):
-                lang_subs = []
+
+            def process_language(container, base_url, lang_code, sub_name, query):
+                lang_subs = container.setdefault(lang_code, [])
                 for fmt in self._SUBTITLE_FORMATS:
                     query.update({
                         'fmt': fmt,
@@ -2324,30 +2325,50 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     lang_subs.append({
                         'ext': fmt,
                         'url': update_url_query(base_url, query),
+                        'name': sub_name,
                     })
-                container[lang_code] = lang_subs
 
-            subtitles = {}
-            for caption_track in (pctr.get('captionTracks') or []):
+            def get_lang_code(track):
+                return (remove_start(track.get('vssId') or '', '.').replace('.', '-')
+                        or track.get('languageCode'))
+
+            def get_text(x):
+                return try_get(x, (lambda x: x['simpleText'],
+                                   lambda x: x['runs'][0]['text']), compat_str)
+
+            subtitles, automatic_captions = {}, {}
+            for lang_code, caption_track in dict(
+                    (get_lang_code(sub), sub) for sub in (pctr.get('captionTracks') or [])).items():
                 base_url = caption_track.get('baseUrl')
                 if not base_url:
                     continue
+                lang_name = get_text(caption_track) or lang_code
                 if caption_track.get('kind') != 'asr':
-                    lang_code = caption_track.get('languageCode')
                     if not lang_code:
                         continue
                     process_language(
-                        subtitles, base_url, lang_code, {})
-                    continue
-                automatic_captions = {}
-                for translation_language in (pctr.get('translationLanguages') or []):
-                    translation_language_code = translation_language.get('languageCode')
-                    if not translation_language_code:
+                        subtitles, base_url, lang_code, lang_name, {})
+                    if not caption_track.get('isTranslatable'):
                         continue
-                    process_language(
-                        automatic_captions, base_url, translation_language_code,
-                        {'tlang': translation_language_code})
-                info['automatic_captions'] = automatic_captions
+                for trans_code, trans_name in dict(
+                        (lang.get('languageCode'), get_text(lang.get('languageName')))
+                        for lang in pctr.get('translationLanguages') or []).items():
+                    if not trans_code:
+                        continue
+                    if caption_track.get('kind') != 'asr':
+                        trans_code += '-' + lang_code
+                        trans_name += ' from ' + lang_name
+                    # Add "-orig" label to the original language so that it can be distinguished
+                    # The subs are returned without "-orig" as well for compatibility
+                    if lang_code == 'a-' + trans_code:
+                        process_language(
+                            automatic_captions, base_url, trans_code + '-orig', trans_name + ' (Original)', {})
+                    # Setting tlang=lang returns damaged subtitles.
+                    # Not using lang_code == f'a-{trans_code}' here for future-proofing
+                    orig_lang = parse_qs(base_url).get('lang', [None])[-1]
+                    process_language(automatic_captions, base_url, trans_code, trans_name,
+                                     {} if orig_lang == trans_code else {'tlang': trans_code})
+            info['automatic_captions'] = automatic_captions
             info['subtitles'] = subtitles
 
         parsed_url = compat_urllib_parse_urlparse(url)
@@ -3427,6 +3448,38 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
             r'\bID_TOKEN["\']\s*:\s*["\'](.+?)["\']', webpage,
             'identity token', default=None)
 
+    def _process_clip(self, url, data):
+        renderer = try_get(
+            data,
+            lambda x: next(
+                (r for r in (s.get('engagementPanelSectionListRenderer')
+                             for s in x['engagementPanels'])
+                 if (r or {}).get('panelIdentifier') == 'engagement-panel-clip-view'),
+                None), dict)
+        if not renderer:
+            return
+        clip_commands = try_get(
+            renderer,
+            lambda x: x['content']['clipSectionRenderer']['contents'][0]['clipAttributionRenderer']['onScrubExit']['commandExecutorCommand']['commands'],
+            list) or []
+        clip_open = try_get(
+            clip_commands,
+            lambda x: next((c['openPopupAction'] for c in x if c.get('openPopupAction')), None),
+            dict)
+        clip_open_commands = try_get(
+            clip_open,
+            lambda x: x['popup']['notificationActionRenderer']['actionButton']['buttonRenderer']['command']['commandExecutorCommand']['commands'],
+            list) or []
+        clip_data = try_get(
+            clip_open_commands,
+            lambda x: next((c['loopCommand'] for c in x if c.get('loopCommand')), None),
+            dict)
+        if clip_data.get('postId') in url:
+            clip_times = (int_or_none(clip_data.get('startTimeMs')),
+                          int_or_none(clip_data.get('endTimeMs')))
+            if None not in clip_times:
+                self._downloader.params['_clip_args'] = '%gms,%gms' % clip_times
+
     def _real_extract(self, url):
         item_id = self._match_id(url)
         url = update_url(url, netloc='www.youtube.com')
@@ -3455,6 +3508,8 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor):
             data, lambda x: x['currentVideoEndpoint']['watchEndpoint']['videoId'],
             compat_str) or video_id
         if video_id:
+            if item_id == 'clip':
+                self._process_clip(url, data)
             return self.url_result(video_id, ie=YoutubeIE.ie_key(), video_id=video_id)
         # Capture and output alerts
         alert = self._extract_alert(data)

youtube_dl/jsinterp.py

@@ -14,6 +14,7 @@ from .utils import (
     remove_quotes,
     unified_timestamp,
     variadic,
+    write_string,
 )
 from .compat import (
     compat_basestring,
@@ -53,15 +54,16 @@ def wraps_op(op):
 
 # NB In principle NaN cannot be checked by membership.
 # Here all NaN values are actually this one, so _NaN is _NaN,
-# although _NaN != _NaN.
+# although _NaN != _NaN. Ditto Infinity.
 
 _NaN = float('nan')
+_Infinity = float('inf')
 
 
 def _js_bit_op(op):
 
     def zeroise(x):
-        return 0 if x in (None, JS_Undefined, _NaN) else x
+        return 0 if x in (None, JS_Undefined, _NaN, _Infinity) else x
 
     @wraps_op(op)
     def wrapped(a, b):
@@ -84,7 +86,7 @@ def _js_arith_op(op):
 def _js_div(a, b):
     if JS_Undefined in (a, b) or not (a or b):
         return _NaN
-    return operator.truediv(a or 0, b) if b else float('inf')
+    return operator.truediv(a or 0, b) if b else _Infinity
 
 
 def _js_mod(a, b):
@@ -220,6 +222,42 @@ class LocalNameSpace(ChainMap):
         return 'LocalNameSpace%s' % (self.maps, )
 
 
+class Debugger(object):
+    ENABLED = False
+
+    @staticmethod
+    def write(*args, **kwargs):
+        level = kwargs.get('level', 100)
+
+        def truncate_string(s, left, right=0):
+            if s is None or len(s) <= left + right:
+                return s
+            return '...'.join((s[:left - 3], s[-right:] if right else ''))
+
+        write_string('[debug] JS: {0}{1}\n'.format(
+            '  ' * (100 - level),
+            ' '.join(truncate_string(compat_str(x), 50, 50) for x in args)))
+
+    @classmethod
+    def wrap_interpreter(cls, f):
+        def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs):
+            if cls.ENABLED and stmt.strip():
+                cls.write(stmt, level=allow_recursion)
+            try:
+                ret, should_ret = f(self, stmt, local_vars, allow_recursion, *args, **kwargs)
+            except Exception as e:
+                if cls.ENABLED:
+                    if isinstance(e, ExtractorError):
+                        e = e.orig_msg
+                    cls.write('=> Raises:', e, '<-|', stmt, level=allow_recursion)
+                raise
+            if cls.ENABLED and stmt.strip():
+                if should_ret or not repr(ret) == stmt:
+                    cls.write(['->', '=>'][should_ret], repr(ret), '<-|', stmt, level=allow_recursion)
+            return ret, should_ret
+        return interpret_statement
+
+
 class JSInterpreter(object):
     __named_object_counter = 0
 
@@ -307,8 +345,7 @@ class JSInterpreter(object):
     def __op_chars(cls):
         op_chars = set(';,[')
         for op in cls._all_operators():
-            for c in op[0]:
-                op_chars.add(c)
+            op_chars.update(op[0])
         return op_chars
 
     def _named_object(self, namespace, obj):
@@ -326,9 +363,8 @@ class JSInterpreter(object):
         # collections.Counter() is ~10% slower in both 2.7 and 3.9
         counters = dict((k, 0) for k in _MATCHING_PARENS.values())
         start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1
-        in_quote, escaping, skipping = None, False, 0
-        after_op, in_regex_char_group = True, False
-
+        in_quote, escaping, after_op, in_regex_char_group = None, False, True, False
+        skipping = 0
         for idx, char in enumerate(expr):
             paren_delta = 0
             if not in_quote:
@@ -382,10 +418,12 @@ class JSInterpreter(object):
         return separated[0][1:].strip(), separated[1].strip()
 
     @staticmethod
-    def _all_operators():
-        return itertools.chain(
-            # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
-            _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS)
+    def _all_operators(_cached=[]):
+        if not _cached:
+            _cached.extend(itertools.chain(
+                # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
+                _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS))
+        return _cached
 
     def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion):
         if op in ('||', '&&'):
@@ -416,7 +454,7 @@ class JSInterpreter(object):
         except Exception as e:
             if allow_undefined:
                 return JS_Undefined
-            raise self.Exception('Cannot get index {idx:.100}'.format(**locals()), expr=repr(obj), cause=e)
+            raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e)
 
     def _dump(self, obj, namespace):
         try:
@@ -438,6 +476,7 @@ class JSInterpreter(object):
     _FINALLY_RE = re.compile(r'finally\s*\{')
     _SWITCH_RE = re.compile(r'switch\s*\(')
 
+    @Debugger.wrap_interpreter
     def interpret_statement(self, stmt, local_vars, allow_recursion=100):
         if allow_recursion < 0:
             raise self.Exception('Recursion limit reached')
@@ -511,7 +550,6 @@ class JSInterpreter(object):
                 expr = self._dump(inner, local_vars) + outer
 
         if expr.startswith('('):
-
             m = re.match(r'\((?P<d>[a-z])%(?P<e>[a-z])\.length\+(?P=e)\.length\)%(?P=e)\.length', expr)
             if m:
                 # short-cut eval of frequently used `(d%e.length+e.length)%e.length`, worth ~6% on `pytest -k test_nsig`
@@ -693,7 +731,7 @@ class JSInterpreter(object):
                 (?P<op>{_OPERATOR_RE})?
                 =(?!=)(?P<expr>.*)$
             )|(?P<return>
-                (?!if|return|true|false|null|undefined)(?P<name>{_NAME_RE})$
+                (?!if|return|true|false|null|undefined|NaN|Infinity)(?P<name>{_NAME_RE})$
             )|(?P<indexing>
                 (?P<in>{_NAME_RE})\[(?P<idx>.+)\]$
             )|(?P<attribute>
@@ -727,11 +765,12 @@ class JSInterpreter(object):
             raise JS_Break()
         elif expr == 'continue':
             raise JS_Continue()
-
         elif expr == 'undefined':
             return JS_Undefined, should_return
         elif expr == 'NaN':
             return _NaN, should_return
+        elif expr == 'Infinity':
+            return _Infinity, should_return
 
         elif md.get('return'):
             return local_vars[m.group('name')], should_return
@@ -760,18 +799,28 @@ class JSInterpreter(object):
             right_expr = separated.pop()
             # handle operators that are both unary and binary, minimal BODMAS
             if op in ('+', '-'):
+                # simplify/adjust consecutive instances of these operators
                 undone = 0
                 while len(separated) > 1 and not separated[-1].strip():
                     undone += 1
                     separated.pop()
                 if op == '-' and undone % 2 != 0:
                     right_expr = op + right_expr
+                elif op == '+':
+                    while len(separated) > 1 and separated[-1].strip() in self.OP_CHARS:
+                        right_expr = separated.pop() + right_expr
+                # hanging op at end of left => unary + (strip) or - (push right)
                 left_val = separated[-1]
                 for dm_op in ('*', '%', '/', '**'):
                     bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim))
                     if len(bodmas) > 1 and not bodmas[-1].strip():
                         expr = op.join(separated) + op + right_expr
-                        right_expr = None
+                        if len(separated) > 1:
+                            separated.pop()
+                            right_expr = op.join((left_val, right_expr))
+                        else:
+                            separated = [op.join((left_val, right_expr))]
+                            right_expr = None
                         break
                 if right_expr is None:
                     continue
@@ -797,6 +846,8 @@ class JSInterpreter(object):
 
             def eval_method():
                 if (variable, member) == ('console', 'debug'):
+                    if Debugger.ENABLED:
+                        Debugger.write(self.interpret_expression('[{}]'.format(arg_str), local_vars, allow_recursion))
                     return
                 types = {
                     'String': compat_str,

youtube_dl/utils.py

@@ -2406,7 +2406,7 @@ class ExtractorError(YoutubeDLError):
         """ tb, if given, is the original traceback (so that it can be printed out).
         If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
         """
-
+        self.orig_msg = msg
         if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
             expected = True
         if video_id is not None: